The invention relates generally to covalent modification of molecules useful in, for example, construction of chemical libraries (including peptide libraries,), modification of surfaces (including cell surfaces), and modification of molecules under physiological conditions (e.g., in a cellular environment).
Nature executes a remarkable feat of achieving molecular recognition in the complex environment of the cellular world. Antibodies bind a single antigen with high affinity, enzymes act on specific substrates in the presence of an enormous variety of similar compounds, and signaling molecules only trigger responses in their target cells. Recently, chemists have attempted to mimic this selectivity by developing a repertoire of reactions that can take place in the presence of other functional groups, and ultimately within the medium of a living cell. Ideally, the reactive partners would be abioitc, react rapidly in water at physiological pH and temperature, form a stable adduct under physiological conditions, and recognize only each other while ignoring their cellular surroundings. The demands on selectivity imposed by cells preclude the use of most conventional covalent reactions. Given these constraints and demands it is not surprising that only a handful of such reactions exist.
When the reaction results in coupling of the two reactants it is termed a xe2x80x9cchemoselective ligationxe2x80x9d (Lemieux et al. Trends Biotechnol. 1998, 16, 506). First described in the arena of protein chemistry, the term is used to describe the coupling of two functional groups in an aqueous environment. The coupling partners are mutually and uniquely reactive, thereby eliminating the need for protecting groups on surrounding functional groups. Chemoselective ligation reactions have been designed for modification of cell surfaces, as well as to provide for ligation reactions in peptide synthesis. Chemoselective ligation reactions have also been designed to modify only one cellular component among all others have provided unique insight into cellular processes (Winans et al. Chem. Biol. 1998, 5, R313).
Three common examples of chemoselective ligation reactions are shown below in Table 1 (Lemieux et al. Trends Biotechnol. 1998, 16, 506). These electrophile-nucleophile pairs have orthogonal reactivity to other functional groups present in many biomolecules.
The chemoselective ligation reaction between a ketone and an aminoooxy or hydrazide group has enabled the engineering of the composition of cell surfaces, and has been used for both in vitro and in vivo chemoselective ligations. Mahal et al. Science 1997, 276, 1125; Brownlee et al. Science 1986, 232, 1629. While not entirely abiotic, the ketone is generally orthogonal to the reactivity of the functional groups present in the outer coating of the cell, which is composed of a variety of heterogeneous glycoproteins) Gumbiner Cell 1996, 84, 345. In vivo chemoselective ligation on the cell surface can be accomplished through unnatural sialic acid biosynthesis. (Kayser et al. J. Biol. Chem. 1992, 267, 16934; Kosa et al. Biochem. Biophys. Res. Comm. 1993, 190, 914; Keppler et al. J. Biol. Chem. 1995, 270, 1308).
Human cells metabolize the unnatural precursor N-levulinoylmannosamine (ManLev, 3 below), a ketone-bearing analog of the native sugar N-acetylmannosamine (1). The substrate promiscuity of this pathway permits the metabolism of the unnatural ManLev precursor 3 into sialic acid analogs 4 on living cells, resulting in the display of ketones (shown as boxes in the schematic below) on the cell surface. These metabolically installed ketones give the cell a unique reactivity, thereby allowing modification of cell surfaces by chemoselective ligation with any moiety bearing a hydrazide or aminooxy group. (Mahal et al. Science 1997, 276, 1125). The ability to perform orthogonal chemical reactions on cell surfaces has enabled the decoration of cells with synthetic glycans (Yarema et al. J. Biol. Chem. 1998, 273, 31168), targeting of MRI probes to tumor cells (Lemieux et al. J. Am. Chem. Soc. 1999, 121, 4278), and production of novel receptors for facilitating viral-mediated gene transfer (Lee et al. J. Biol. Chem. 1999, 274, 21878). 
While useful for cell surface chemistry, ketone ligation reactions have limited intracellular utility due to competition with endogenous keto-metabolites. Tsien and coworkers reported a second chemoselective ligation reaction that circumvents this problem by providing for condensation of a unique cysteine-rich hexapeptide motif with a bis-dithioarsolane (Griffin et al. Science 1998, 281, 269). This enabled the targeting of a synthetic fluorescent dye to a single protein within the environs of a living cell.
In addition to its usefulness for modification of cell surfaces, the unique reactivity of the ketone has also been exploited for glycopeptide synthesis. Ketone-bearing amino acids have been incorporated into a synthetic peptide, allowing subsequent chemoselective ligation. This method was used to glycosylate synthetic peptides at defined locations via an unnatural oxime linkage (illustrated in the schematic below), enabling the synthesis of otherwise intractable glycoproteins and permitting the investigation of the effect of glycosylation on protein structure and function (Marcaurelle et al. Tetrahedron Lett. 1998, 39, 8417); (Marcaurelle et al. Tetrahedron Lett. 1998, 39, 7279); (Rodriguez et al. J. Org. Chem. 1998, 63, 7134). 
A subset of chemoselective ligations results in the formation of a native bond within a peptide. Kent and coworkers developed such a method, termed native chemical ligation, for the construction of large proteins that are far beyond the realm of stepwise solid phase peptide synthesis. (Canne et al. J. Am. Chem. Soc. 1996, 118, 5891). The reaction exploits a selective trans-thioesterification (as depicted in the reaction below). One peptide segment bearing a C-terminal thioester reacts with another bearing an N-terminal cysteine residue. The fast initial step is followed by a spontaneous irreversible rearrangement to form a native peptide bond. This methodology has been exploited for the synthesis of proteins containing suitably positioned cysteine residues. (Dawson et al. Science 1994, 266, 776). 
The addition of new chemoselective reactions to the rather limited existing panel would expand the utility of this chemistry and enable novel applications. For example, the cell surface display of two different, orthogonally reactive functional groups would allow the tandem delivery of biological probes, drugs or homogeneous carbohydrate moieties in a chemically controlled manner. Furthermore, an entirely abiotic functional group could be used to form covalent adducts within a cell, thus providing a unique target for modification. In addition, a novel reactive pair would enable native chemical ligation at any site in a peptide backbone, not just at cysteine residues.
The Staudinger reaction, which involves reaction between trivalent phosphorous compounds and organic azides (Staudinger et al. Helv. Chim. Acta 1919, 2, 635), has been used for a multitude of applications. (Gololobov et al. Tetrahedron 1980, 37, 437); (Gololobov et al. Tetrahedron 1992, 48, 1353). There are almost no restrictions on the nature of the two reactants. The phosphines can be cyclic or acyclic, halogenated, bisphosphorus, or even polymeric. Similarly, the azides can be alkyl, aryl, acyl or phosphoryl. The one restriction on all of the reactions mentioned thus far is that they were carried out, or at least initiated, under oxygen-free anhydrous conditions since most phosphorus (III) compounds are readily oxidized and many are susceptible to hydrolysis.
The mechanism of the reaction has been studied in detail (Leffler et al. J. Am. Chem. Soc. 1967, 89, 5235). As illustrated below, the first intermediate is the adduct formed by the attack of the phosphorus lone pair on the terminal nitrogen of the azide. The phosphazide decomposes with the loss of N2 to form an aza-ylide via a 4-membered ring transition state. The aza-ylide can be isolated if the substituents are able to stabilize the highly nucleophilic nitrogen atom and electrophilic phosphorus atom. However, when water is added to the reaction mixture the aza-ylide rapidly hydrolyzes to form a phosphine oxide and an amine. Depending on the concentration of the reactants in solution and the electronic and steric nature of the substituents, the rate determining step can be either the association to form the phosphazide, or its breakdown to the aza-ylide (Gololobov et al. Tetrahedron 1992, 48, 1353) 
Traditionally the Staudinger reaction has simply entailed the reduction of azide functionalities to amines by triphenylphosphine initiated under anhydrous conditions. However, several reported reactions take advantage of the reactivity of the aza-ylide to carry out chemical transformations besides hydrolysis to the amine, some even in the presence of water. When the phosphorus compound possesses at least one alkoxy substituent, the aza-ylide undergoes an Arbuzov-like rearrangement resulting in a new covalent linkage, as shown in reaction (a) below (Keogh et al. J. Org. Chem. 1986, 51, 2767). Triphenylphosphine can be used to mediate the formation of amide bonds without the need for an activated ester (see reaction (b) below). (Wilt et al. J. Org. Chem. 1985, 50, 2601). In addition, an aza-ylide can be trapped by an intramolecular electrophile such as an alkyl halide ((c) below), (Mastryukova et al. Zh. Obshch. Khim. 1988, 58, 1967) or ester ((d) below), (Khoukhi et al. Tetrahedron Lett. 1986, 27, 1031). 
Prior to the present disclosure, the Staudinger reaction has never been adapted to perform ligations in a biological environment. Both reactants are intrinsically orthogonal to biological molecules and yet the azide is readily installed in carbohydrates or proteins. The reaction between the phosphine and the azide must produce a stable covalent adduct and utilize a phosphorus compound that is stable to water and air. Reactions (a) and (d) above appear to fulfill this latter requirement as they are carried out in the presence of a small amount of water, however it was unknown whether this reactivity would extend to truly aqueous conditions.
There is a need in the field for additional mechanisms to modify biological molecules through chemoselective ligations, particularly in a biological environment. The present invention addresses this need.
The present invention features a chemoselective ligation reaction that can be carried out under physiological conditions. In general, the invention involves condensation of a specifically engineered phosphine, which can provide for formation of an amide bond between the two reactive partners resulting in a final product comprising a phosphine oxide, or which can be engineered to comprise a cleavable linker so that a substituent of the phosphine is transferred to the azide, releasing an oxidized phosphine byproduct and producing a native amide bond in the final product. The selectivity of the reaction and its compatibility with aqueous environments provides for its application in vivo (e.g., on the cell surface or intracellularly) and in vitro (e.g., synthesis of peptides and other polymers, production of modified (e.g., labeled) amino acids).